Searcharxiv⌕ Search

arXiv subjects

Junjie Zhang

Publications and source records attributed to Junjie Zhang.

At least 19 recordsLinked to original sources
arXiv2026

Ultrafast Magnetization Induced from Raman-Active Axial Chiral Phonons

Axial chiral phonons, which carry angular momentum and exhibit unusually large magnetic moments, provide a new degree of freedom for controlling the time-reversal symmetry and magnetic properties of quantum materials. While infrared-active axial phonons are efficiently excited to large amplitudes by circularly polarized terahertz pulses, their application may be constrained by short lifetimes, small penetration depths, low spatial resolution, and limited availability of optical sources. Here, we report that an axial Raman-active phonon mode in CeF3 exhibits a significant magnetic moment and an exceedingly long lifetime that more closely matches the paramagnetic spin dynamics. The axial phonon population is resonantly driven by a near-infrared laser pulse with rotating linear polarization, known as an optical centrifuge, a coherent control scheme never applied to solids before. The phonon-driven magnetization observed by time-resolved Faraday rotation scales quadratically with incident power and rapidly decreases at high temperatures, consistent with phonon inverse Faraday effect from many-body spin-phonon coupling. Our findings open a new avenue for using shaped laser pulses from widely accessible light sources to manipulate coherent axial chiral phonons and ultrafast spintronics.

cond-mat.mtrl-sci↗
arXiv2026

RedEvoAgent: Automatic Red-Teaming Agent with Experience-Driven Skill Evolution

LLM-based agents are increasingly deployed in product-level execution harnesses, where jailbreaks can trigger harmful tool use and persistent state changes, creating greater risks than unsafe text generation alone. Existing automatic red-teaming methods often rely on fixed attacks, while recent agentic attackers coordinate multiple jailbreak tools and show stronger potential through trajectory-based retrieval. However, such retrieval can reuse misleading experiences due to retrieval bias and unclear tool credit, and full trajectories add context overhead while reducing interpretability. We propose RedEvoAgent, a black-box red-teaming agent that distills cross-case attack trajectories into a concise, human-readable attack skill. The attack skill adaptively evolves through tool-effectiveness profiling and Deciding-Tool Attribution for skill updates, and a validation ratchet that retains only updates improving validation performance. Experiments on multiple benchmarks, target models, and target execution harnesses show that RedEvoAgent outperforms fixed and agentic baselines, improves tool efficiency, and transfers across attacker models and target execution harnesses.

cs.CR↗
arXiv2026

Construction of $n$-cotorsion pairs in some abelian categories

In this paper, we mainly investigate $n$-cotorsion pairs in two types of abelian categories. Firstly, we construct left (right) $n$-cotorsion pairs in comma categories by improving some isomorphisms between homology groups and combining some classes of objects in comma categories. Secondly, based on known homological formulas over trivial ring extensions, we study some classes of modules satisfying certain conditions over such rings, and we establish left (right) $n$-cotorsion pairs over trivial ring extensions through these classes. Then corresponding conclusions over Morita rings with zero bimodule homomorphisms are directly presented. The hereditary property of left (right) $n$-cotorsion pairs is investigated in both cases. Finally, we apply main results obtained in comma categories and over trivial ring extensions to formal triangular matrix rings. At the same time, via these applications, we found that relevant conditions in known results have been improved and the known results have been elevated.

math.RA↗
arXiv2026

G2VD: Generalizable AI-Generated Video Detection via Counterfactual Intervention and Causal Disentanglement

Rapid advances in AI video generation pose increasing security risks and call for reliable detectors with strong cross-domain generalization. Although existing methods perform well under in-domain evaluation, their performance degrades substantially on unseen generators. A key reason is shortcut learning, where detectors rely on domain-specific bias rather than intrinsic forensic cues. To address this issue, we propose G2VD, a generalizable AI-generated video detection framework based on counterfactual intervention and causal disentanglement. First, G2VD introduces a counterfactual intervention pipeline (CFIPipeline) that constructs counterfactual samples through VAE-based reconstruction and subsequent frequency-domain and pixel-domain alignment, thereby weakening spurious correlations between domain-specific bias and authenticity labels. Building on this intervention, we further design a causal disentanglement classifier that combines two domain-anchored branches with complementary objectives and a constraint based on the Hilbert-Schmidt Independence Criterion (HSIC), encouraging the causal and non-causal representations to capture intrinsic forensic cues and domain-specific bias, respectively. Experiments across four public datasets demonstrate strong cross-domain performance and consistent gains over baseline methods. In the challenging GenVidBench setting, G2VD achieves over 90\% overall ACC, with improvements of 0.194 in F1 and 0.104 in AUC over comparable state-of-the-art methods, while using only 10\% of the available training data. Code is available at https://github.com/DMOSCAR-98/G2VD.

cs.CV↗
arXiv2026

MathCoPilot: An Interactive System for Human-AI Symbiotic Paradigm of Mathematical Research

Existing LLM-based theorem provers have achieved impressive results on formal mathematics benchmarks, yet they remain confined to acting as autonomous agents that prove a stated proposition. In this paper, we propose MathCoPilot, a human-in-the-loop system that embodies a new human--AI symbiotic paradigm for mathematical research, in which the mathematician steers the high-level mathematical direction while AI agents carry out the detailed formalization and proof work under continuous human guidance. MathCoPilot unifies three core capabilities: (1) an interactive workbench where the mathematician and AI agents collaborate through a living proof blueprint that decomposes a proof into navigable steps the human can directly inspect, direct, and refine; (2) automated proving skill orchestration with adaptive knowledge base search and Lean-integrated iterative verification; and (3) topic-driven paper retrieval and automated formalization into a verified Lean knowledge base. Using MathCoPilot, we systematically compare four state-of-the-art LLMs, including Gemini~3.1~Pro, GPT-5.4, and Claude~Opus~4.7, on a FormalMATH subset and on two real PDE theorems requiring deep domain expertise, evaluating their ability to produce verified Lean~4 proofs and to identify errors in deliberately incorrect proofs. Our results show that while current models can handle undergraduate-level problems with high success rates under favorable autoformalization conditions, substantial challenges remain for domain-specific theorems requiring genuine mathematical understanding.

cs.AI↗
arXiv2026

Hidden Decoding at Scale: Latent Computation Scaling for Large Language Models

Scaling Large Language Models (LLMs) has been driven mainly by enlarging the Transformer backbone, but for an already-strong model this requires another round of costly pretraining. We study whether an existing backbone can keep improving by allocating more computation to each token while leaving the Transformer backbone fixed. Depth-recurrent (looped) Transformers pursue this goal but are hard to scale, because looped computation does not fit naturally with the pipeline parallelism used to train the largest models. We add computation along the sequence-length dimension, where the extra computation is simply a longer input and stays compatible with standard large-model training. We propose Hidden Decoding, a sequence-length scaling method applied during continued pretraining (CPT). It expands each token into n streams with independent embedding tables and keeps the intermediate streams' key-value cache as context, so each token performs more internal computation without adding or widening Transformer layers. To keep this affordable at scale, we introduce Stream-Factorized Attention, in which most layers attend only within each stream and only a few layers mix across streams, reducing the attention cost from quadratic to roughly linear in n. Experiments support two scaling results. At frontier scale, we train WeLM-HD4-80B and WeLM-HD4-617B at n=4 and improve their matched non-HD baselines, making Hidden Decoding the first demonstrated sequence-length scaling method at the 100B+ MoE scale. Across expansion factors, the gains grow as n increases, showing that sequence-length expansion is a practical fixed-backbone scaling path for frontier-scale LLMs.

cs.CL↗
arXiv2026

Entropy Pacing Policy Optimization for Multi-Task Agentic Reinforcement Learning

Recent breakthroughs of Reinforcement Learning (RL) have highlighted its potential for complex agentic Large Language Model (LLM) tasks. However, existing efforts largely focus on single-task settings, whereas real-world deployment necessitates a generalist agent capable of solving multiple tasks simultaneously. In this work, we identify a critical yet underexplored phenomenon in multi-task agentic RL: different tasks can exhibit exploration-exploitation pace mismatch. Specifically, easier tasks may converge early to low-entropy policies that hinder learning on harder tasks, while harder tasks can, in turn, push easier tasks back toward high-entropy exploration. This back-and-forth interaction creates inter-task entropy crossovers and frequent entropy spikes. Inspired by this observation, we introduce Entropy Pacing Policy Optimization (EPPO) for multi-task agentic LLMs, which coordinates entropy across tasks to stabilize multi-task optimization. At the core of EPPO is a task-wise dynamic clipping mechanism that replaces the fixed clipping threshold in Group Relative Policy Optimization (GRPO) with a task entropy-aware adaptive bound, tightening updates for over-confident tasks while relaxing them for under-explored ones. Experiments on the multi-task agentic benchmarks demonstrate that the proposed EPPO yields results superior to its counterparts.

cs.LG↗
arXiv2026

Large-Scale Tunnel Air-Ground Collaboration With FLISP: Fast LiDAR-IMU Synchronized Path Planner

Hydropower tunnel inspection is critical for infrastructure integrity yet remains inefficient and hazardous using manual methods. We propose FLISP (Fast LiDAR-IMU Synchronized Path Planner), a mapless planning framework for cooperative UGV-UAV inspection. Unlike traditional map-based paradigms, FLISP features three core contributions: (1) a unified architecture where a single UGV-mounted LiDAR-IMU suite drives synchronized path generation for both platforms; (2) platform-specific solvers utilizing an enhanced Firefly Algorithm for UGV obstacle avoidance and a dynamic iterative optimizer for UAV flight; and (3) a hierarchical refinement strategy ensuring kinematic feasibility without state estimation drift. Benchmarks in a 1.2 km operational tunnel demonstrate that FLISP circumvents structural bottlenecks of map-based methods, eliminating map rasterization overhead (Fast-LIO2 + A*) and sampling instability (LIO-SAM + RRT*). FLISP achieves a 100% success rate with 7 ms latency, representing a 7-fold speedup over grid-based and a three-order-of-magnitude improvement over sampling-based baselines. Validated in operational hydropower tunnels, this approach offers a scalable solution for robotic inspection in feature-degraded linear infrastructure. A demonstration video is available at https://youtu.be/Y_ezs1PfLJ4, and the code at https://github.com/ArchibaldGuo/FLISP.git.

cs.RO↗
arXiv2026

Ultra-Fusion: A Resilient Tightly-Coupled Multi-Sensor Fusion SLAM Framework under Sensor Degradation and Spatiotemporal Perturbation for Intelligent Transportation Systems

Reliable localization is essential for intelligent transportation systems (ITS), including autonomous vehicles, quadruped last-mile carriers, and infrastructure-inspection unmanned aerial vehicles (UAVs). Although tightly-coupled multi-sensor fusion improves accuracy in favorable conditions, deployed systems remain vulnerable to sensor degradation -- poor illumination, LiDAR degeneracy, wheel slippage, and GNSS outage -- and to spatiotemporal calibration errors. These failures are common in urban canyons, tunnels, and high-speed corridors, where localization drift can degrade route tracking, tunnel passage continuity, and local map alignment. This paper presents Ultra-Fusion, a tightly-coupled multi-sensor localization framework based on a unified sliding-window estimator. Asynchronous measurements are timestamp-ordered and converted into optional factors within one optimization window, supporting WIO, VIO, LIO, and LVIO with optional wheel and GNSS augmentation. Observability-aware initialization selects the bootstrap mode, factor-wise reliability scheduling gates degraded measurements, and online LiDAR--IMU spatiotemporal calibration refines temporal offsets and rotational extrinsics during operation. We extend the M3DGR benchmark with simulation trajectories and evaluate more than 60 open-source SLAM systems on M3DGR, M2DGR-Plus, KAIST, GrandTour, and MARS-LVIG. The results show competitive accuracy across wheeled, legged, and aerial platforms under long-duration and high-speed operation, degradation, and calibration perturbation, improving localization availability for road-level autonomy, campus and warehouse mobility, and low-altitude aerial inspection. To benefit the industrial and academic community, we will release source code and datasets upon paper acceptance.

cs.RO↗
arXiv2026

Observation of correlated plasmons in low-valence nickelates

The discovery of nickelate superconductors has opened a new arena for studying the behavior of correlated electron liquids that give rise to unconventional superconductivity. While critical information about a material's charge dynamics is encoded in its plasmons, collective modes of the electron gas, these excitations have not yet been observed in nickelate materials. Here, we use resonant inelastic x-ray scattering (RIXS) to detect plasmons in the metallic, low-valence nickelate Pr4Ni3O8. Although qualitatively similar to those in cuprates, the nickelate plasmons are more heavily damped and have a lower velocity than those in a cuprate at comparable doping, which we attribute to reduced electronic hopping and enhanced screening of the long-range Coulomb interactions. Furthermore, the plasmons in Pr4Ni3O8 soften with increasing temperature, in contrast to the cuprate, where plasmons remain at nearly fixed energy but become more strongly damped. Taken together, these results reveal a distinct charge-screening landscape in nickelates and place quantitative constraints on analogies to cuprates.

cond-mat.str-el↗
arXiv2026

Scaling Learning-based AEB with Massive Unlabeled Data

This paper studies how to scale learning-based automatic emergency braking (AEB) with massive unlabeled fleet data under production constraints. Our approach is based on meta-feedback semi-supervised learning (MF-SSL), where a teacher generates pseudo labels for unlabeled driving data and is updated using a small labeled anchor set as safety-critical feedback. In production, anchor ambiguity and labeled-unlabeled mismatch can amplify systematic pseudo-label errors, leading to spurious triggers. We propose a stabilized MF-SSL framework with (i) Noise-Aware Decoupling, which removes ambiguity-prone anchors from the teacher's supervised update path, and (ii) kinematics-gated pseudo-labeling with a teacher conflict penalty to suppress mismatch-induced risk hallucinations on unlabeled data while maintaining broad coverage. Extensive experiments show consistent gains as unlabeled data scale from 1M to 1B windows, improving safety while keeping comfort stable. The 1B-trained student model is deployed to hundreds of thousands of vehicles and validated over \$10^9$ km of driving, achieving a positive-to-false activation ratio exceeding 100:1 and a 35% improvement in accident-free driving mileage over a production rule-only baseline.

cs.LG↗
arXiv2026

Tool-IQA: Augmenting Image Quality Assessment with Simple Tools

Vision-Language Models (VLMs) have been increasingly adopted for Image Quality Assessment (IQA). However, current methods typically employ a static one-shot scoring paradigm, despite the fact that humans assess image quality through dynamic visual inspection, e.g., selectively adjusting views to verify details and subtle artifacts. Specifically, relying solely on a single-pass observation introduces two primary limitations: first, perceiving the image only at a global scale restricts the assessment of finer local details; second, the original intensity distribution of the image may overwhelm the visibility, leading to insufficient inspection of image quality. To address these issues, we propose Tool-IQA, shifting the assessment mechanism from passive scoring to a tool-augmented workflow. In particular, we equip VLMs with simple yet effective view tools: a Magnifier to inspect local details, and a Gamma Corrector to uncover visibility and hidden artifacts. The assessment follows a structured pipeline that consists of an initial observation with rubric notes, a tool-augmented in-depth inspection, and a final quantification for calibrated quality score. Furthermore, to ensure efficient and purposeful tool callings, we introduce a batch-aware training strategy to reward tool interactions that can yield positive contributions rather than simply encouraging usage. Experiments on a variety of IQA benchmarks demonstrate that, with effective tool calling and calibrated assessment, our proposed Tool-IQA significantly outperforms existing state-of-the-art models, e.g., it achieves a PLCC of 0.854 on the challenging CLIVE dataset.

cs.CV↗
arXiv2026

Solved in Unit Domain: JacobiNet for Differentiable Coordinate-Transformed PINNs

Physics-Informed Neural Networks (PINNs) offer a powerful framework for solving PDEs by embedding physical laws into the learning process. However, when applied to domains with irregular boundaries, PINNs often suffer from instability and slow convergence, which stems from (1) inconsistent normalization due to geometric anisotropy, (2) inaccurate boundary enforcement, and (3) imbalanced loss term competition. A common workaround is to map the domain to a regular space. Yet, conventional mapping methods rely on case-specific meshes, define Jacobians at pre-specified fixed nodes, reformulate PDEs via the chain rule-making them incompatible with modern automatic differentiation, tensor-based frameworks. To bridge this gap, we propose JacobiNet, a learning-based coordinate-transformed PINN framework that unifies domain mapping and PDE solving within an end-to-end differentiable architecture. JacobiNet enables direct Jacobian computation via autograd, shares computation graph with downstream PINNs, thereby avoiding case-specific meshing, explicit Jacobian computation/storage, and manual PDE reformulation while unlocking geometric-editing operations. Separating physical modeling from geometric complexity, JacobiNet (1) addresses normalization challenges in the original anisotropic coordinates, (2) facilitates the hard enforcement of boundary conditions, and (3) mitigates the long-standing imbalance among loss terms. Evaluated on various PDEs, JacobiNet reduces the relative L2 error from 0.11-0.73 to 0.01-0.09, achieving an average 15.6x improvement in accuracy. In vessel-like domains with varying shapes, JacobiNet enables millisecond-level mapping inference for unseen geometries, improves prediction accuracy by an average of 3.65x, while delivering over 10x speedup-demonstrating strong generalization, accuracy, and efficiency.

cs.LG↗
arXiv2026

Unveiling the Reasoning Process of Large Language Models

Large language models often reason beyond surface tokens, but the internal stage at which token-level information becomes abstract relational structure remains unclear. We investigate this question by analyzing how attention heads and layers transform information during autoregressive reasoning. Across mathematical and symbolic reasoning tasks, we observe a consistent layer-wise division of labor: outer layers mainly preserve and route input-related features, whereas middle layers reorganize them into more transferable rule-level representations. This interpretation is supported by representation geometry: middle-layer states occupy lower-dimensional manifolds and show stronger alignment across disjoint vocabularies that instantiate the same symbolic rules. It is further supported by causal interventions: removing middle-layer components identified by our interaction-based criterion produces substantially larger downstream changes and accuracy drops than removing components from other regions or at random. Together, these results suggest that abstract reasoning is not uniformly distributed across transformer layers, but is preferentially formed in a middle-layer computation stage that converts token-level information into reusable relational structure.

cs.AI↗
arXiv2026

BatchWeave: A Consistent Object-Store-Native Data Plane for Large Foundation Model Training

Modern Large Foundation Model (LFM) training has transformed the data pipeline from a static ingestion layer into a dynamic component that must co-evolve with the training process. Existing systems are ill-equipped: colocated dataloaders offer no failure isolation, while message queue-based disaggregated dataloaders operate on a record/offset abstraction that cannot express the batch-level semantics required by distributed training. We present BatchWeave, an object-store-native training data plane for distributed LFM training. BatchWeave uses versioned manifests and conditional object writes to coordinate batch publication, recovery, and lifecycle management. First, it introduces the Transactional Global Batch (TGB), which builds on versioned-manifest ACID storage semantics and extends them with training-specific consistency, including atomic all-rank batch visibility, a globally ordered step sequence, checkpoint-aligned lifecycle management, and end-to-end exactly-once recovery. Second, it realizes recovery and retention directly in the storage layer, by durably persisting producer state through the commit protocol and tying reclamation to distributed checkpoint state. Third, its Decentralized Adaptive Commit (DAC) algorithm sustains stable ingestion throughput as the manifest grows, without any inter-producer communication. Evaluations on large-scale multimodal pre-training and SFT workloads using 64 GPUs show that BatchWeave outperforms colocated dataloader throughput while providing full failure isolation, outperforms Apache Kafka in ingestion throughput, and achieves lower consumer read latency than Kafka.

cs.DC↗
arXiv2026

Real-time reinforcement learning for turbulent state-dependent control in a bluff-body wake

Controlling turbulent dynamics remains a major challenge because of its chaotic, multi-scale dynamics, which strongly influence the performance of many fluid systems. Here we report REACT (Reinforcement Learning for Environmental Adaptation and Control of Turbulence), an autonomous reinforcement learning framework for real-time state-dependent control of turbulent wake dynamics in a real wind-tunnel environment. Deployed on an Ahmed-body model equipped solely with onboard sensors and servo-actuated surfaces, REACT learns directly from sparse experimental measurements in a wind-tunnel environment, bypassing empirical turbulence models. The agent autonomously converges to a policy that reduces aerodynamic drag while achieving net energy savings. Without prior knowledge of flow physics, it discovers that dynamically suppressing spatiotemporally coherent flow structures in the bluff-body wake maximizes energy efficiency, achieving two to four times greater performance than model-based baseline controllers. We contrast the state-dependent, dynamics-aware policy of REACT with representative quasi-steady, mean-flow-oriented policies learned by standard reinforcement learning baselines, which deliver lower drag reduction and no direct suppression of coherent instabilities in this turbulent-wake regime. Finally, by training in a nondimensional state-reward space whose amplitudes are approximately Reynolds-number-invariant, and by conditioning on Reynolds number for temporal adaptation, REACT learns a single offline policy that remains effective across the tested Reynolds-number range 86,400 to 518,400, without retraining. These results demonstrate autonomous closed-loop reinforcement learning control in a high-Reynolds-number wind-tunnel environment and suggest a path toward data-driven state-dependent control of turbulent flows.

physics.flu-dyn↗
arXiv2026

STRIDE: Learnable Stepwise Language Feedback for LLM Reasoning

Recent advances in Reinforcement Learning (RL) have underscored its potential for incentivizing reasoning capabilities of Large Language Models (LLMs). However, existing step-level efforts suffer from costly annotations that limit domain coverage, while scalar scores further impose an information bottleneck, offering insufficient semantic bandwidth to improve intermediate decisions. Alternative language-critique approaches, which rely on frozen or external critics, provide richer textual feedback but lack the scalability needed for sustained policy improvement. In this work, we propose language-driven stepwise trajectory redirection, termed as STRIDE, a novel training framework that shifts process supervision from scalar rewards to learnable stepwise language feedback. Specifically, we co-train a generator and a generative verifier using only outcome-based rewards, eliminating external annotations, while delivering sustained policy improvement through jointly aligned verifier training. The verifier's stepwise language critiques explicitly localize and explain failures, enabling the generator to redirect reasoning trajectories at intermediate steps toward alternative decisions. The trajectory redirection design guarantees harmless policy improvement, even under noisy or suboptimal verifier feedback. Experiments on diverse reasoning benchmarks show that STRIDE significantly outperforms state-of-the-art baselines, as well as achieving breakthroughs on zero-pass-rate problems where scalar methods yield no learning signal in our ablation studies, demonstrating the effectiveness of learnable stepwise language feedback for enhancing LLM reasoning.

cs.LG↗
arXiv2026

Bulk superconductivity up to 96 K in pressurized nickelate single crystals

Recently, the Ruddlesden-Popper bilayer nickelate $La_3Ni_2O_7$ has emerged as a superconductor with a transition temperature ($T_c$) of approximately 80 K above 14 GPa (Refs. 1-3). Achieving higher $T_c$ in nickelate superconductors, along with the synthesis of reproducible high-quality single crystals without relying on high-oxygen-pressure growth conditions, remains a significant challenge$^{[4-7]}$. Here we report superconductivity up to 96 K under high pressure in bilayer nickelate single crystals synthesized at ambient pressure. Energy-dispersive spectroscopy, single-crystal X-ray diffraction, nuclear quadrupole resonance and scanning transmission electron microscopy evidenced high crystal quality of the flux-grown $La_2SmNi_2O_{7-δ}$ single crystals. $La_2SmNi_2O_7$ exhibits clear bulk superconductivity, including zero resistivity ($T_{c,max}^{onset}$ = 92 K and $T_{c,max}^{zero}$ = 73 K at 21.6 GPa) and the Meissner effect ($T_c$= 60 K at 20.6 GPa). A low-temperature high-pressure structural study indicates that both monoclinic and tetragonal structures can support superconductivity in this bilayer nickelate. Furthermore, we established a correlation between higher $T_c$ under high pressures and larger in-plane lattice distortion under ambient conditions, corroborated by observing even higher $T_c^{onset}$ of 96 K in $La_{1.57}Sm_{1.43}Ni_2O_{7-δ}$. This study overcomes key limitations in growing nickelate superconductor crystals, resolves the crystal structure in the superconducting state and demonstrates an effective pathway towards achieving higher $T_c$.

cond-mat.supr-con↗